What this estimate does and does not do
This calculator explains downtime-style SLO budgeting. It is suitable for quick SRE planning, incident review, and communicating what an availability target means in minutes. Real production SLOs may be based on request success rate, latency, data freshness, or partial degradation rather than manually entered downtime minutes. Treat the output as a model of your assumptions, not as an authoritative compliance report.
Formula used
An error budget is the amount of unreliability allowed by an SLO over a time window.
allowed downtime = (1 - SLO) × window minutes
remaining budget = allowed downtime - downtime so far
used budget % = downtime so far / allowed downtime
elapsed window % = elapsed days / total window days
burn rate = used budget % / elapsed window %Worked example
For a 99.9% SLO over a 30-day window:
allowed downtime = 0.001 × 30 × 24 × 60 = 43.2 minutesIf you have already used 5 minutes after 7 days, you have used 11.6% of the budget after 23.3% of the window, so your burn rate is below 1×.
How to interpret burn rate
- Below 1×: you are consuming budget slower than calendar time.
- Around 1×: you are on pace to use the budget exactly.
- Above 1×: you are burning budget too quickly.
- Above 2×: investigate before the budget is exhausted.
Assumptions
- This calculator treats downtime minutes as already mapped to the same service, SLO, and measurement window.
- It does not distinguish full outage minutes from partial degradation unless your downtime input already accounts for that.
- Real SLOs often use request success rate, latency, or availability, not just manually entered downtime.
Frequently asked questions
How do you calculate allowed downtime from an SLO?
Allowed downtime equals total time in the window multiplied by one minus the SLO. For example, 99.9% over 30 days allows 0.1% of 43,200 minutes, or 43.2 minutes.
What does burn rate mean?
Burn rate compares how quickly you are consuming the error budget against how much of the window has elapsed. A burn rate above 1 means you are on pace to exhaust the budget early.
Is an error budget always measured in downtime minutes?
No. Real SLOs may be based on request success rate, latency, availability, or another SLI. Downtime minutes are a simple way to explain availability budgets.
What should happen when the budget is exhausted?
Many SRE processes slow or freeze risky releases and shift focus to reliability work when an error budget is exhausted, but the exact policy depends on the team and service criticality.